PD is the reference standard surgical method for the treatment of tumors in the head of the pancreas, the lower end of the common bile duct, the duodenal papilla, and certain benign lesions. The current perioperative mortality of PD is less than 5%[8], but the incidence of perioperative complications is up to 27.1% or even higher[9]. The prognosis of PD is poor, and the 5-year survival rate after surgery is only 16%-20%[10, 11]. Common complications of PD include pancreatic fistula, bile leakage, GI anastomotic leakage, GI bleeding, AKI, gastroparesis, cardiac complications, respiratory complications, UTI, postoperative sepsis, postoperative shock, wound complications, etc.[12, 13] Previous studies have shown that perioperative complications have an adverse effect on survival and prognosis in various tumor types including esophageal cancer[14], gastric cancer[15] and colorectal cancer[16]. Similarly, in pancreatic cancer, the occurrence of perioperative complications also adversely affects survival and prognosis.
AF is the most common type of arrhythmia[4], accounting for 1–2% of the overall population. Patients with AF are at higher risk of thrombo-embolic disease and often require regular anticoagulation therapy[17]. It has been reported that AF can significantly promote the occurrence of perioperative complications of various surgical operations including urinary system surgery and liver transplantation, etc.[6, 7, 18]
To the best of our knowledge, this is the first population-based study that investigated the impact of AF on the outcomes of pancreatic cancer patients with OPD. The reports on mortality of PD varies in various studies. Cameron et al. summarized 1,000 consecutive PD and found that only 10 postoperative deaths, for a mortality of 1%[19]. Kneuertz et al. reported that the 30-day mortality after PD was 2.9%[20]. Our study found that during three-year time from 2012 to 2014, the overall in-hospital mortality was 3.17%. The in-hospital mortality was 7.6% and 2.66% in the AF group and non-AF group, respectively, without significant difference (adjusted OR 3.57, 95% CI 0.86–14.87, P = 0.08). Kumar et al.[21] found that pulmonary complications were the leading cause for mortality after PD. Other causes of death include bile sepsis, liver failure, MI, pancreatic leak, etc. Narayanan et al.[22] showed that the most common cause of death after PD within 90 days is multisystem organ failure, followed by post-pancreatectomy hemorrhage, and cardiopulmonary arrest from myocardial infarction or pulmonary embolus. LOS was longer in AF group, but no significant difference was found (14.6 days versus 11.8, adjusted coefficient 0.16 (0-0.33), P = 0.052). There was a significant difference in cost between the two groups (54601.47 dollars versus 38812.80 dollars, adjusted coefficient 0.35 (0.06–0.64), P = 0.02). The need of more thorough evaluation of cardiac function and more complex management of perioperative complications might be a possible reason.
Our study found that the AF group had more perioperative complications than the non-AF group, including GI anastomotic leakage, cardiac complications, respiratory complications, PE and postoperative shock. Among all complications, the incidence of GI anastomotic leakage in the AF group was approximately 22 times higher than that in the non-AF group. Although AF may affect the stability of hemodynamics and thus affect the blood supply of the anastomosis, the number of occurrences in both groups is small (5 patients versus 30), so further study is needed to confirm the conclusion.
Botwinick et al.[23] reported that patients with AF are more likely to experience gastroparesis, but our study did not find the adverse effect of AF on gastroparesis. There were 249 patients included in Botwinick’s study, of which only 13 had atrial fibrillation, so further research may be needed to confirm the relationship between atrial fibrillation and gastroparesis. In addition, no differences were found in GI bleeding, AKI, gastroparesis, blood transfusion, pneumonia, UTI, cardiac arrest, postoperative sepsis, wound complications, etc.
With the advancement of surgical techniques, more and more elderly pancreatic cancer patients with complex comorbidities receive PD. Several studies have shown that age has an important adverse effect on the occurrence of perioperative complications and survival after PD[24, 25]. So, we further analyzed patients aged > 65 years. No significant difference was found in mortality, while there was significant difference between the two groups in LOS and cost. As for complications, the differences in various complications between the AF group and the non-AF group in the elderly population are similar to the overall population. It can be seen that although age will increase complications and mortality, it will increase evenly between the two groups. Further studies are needed to reveal the influence of age on perioperative complications and death after whipple’s surgery.
There are several limitations of the present study. First of all, our study is a retrospective study based on the NIS database with all inherent shortcomings. Secondly, because NIS is an inpatient database coding with ICD-9-CM, complications without ICD-9 code, such as pancreatic fistula, bile leakage, cannot be analyzed. In addition, tumor information such as tumor type, stage, and neoadjuvant chemotherapy is absent in the NIS database, which may have an impact on the outcomes of OPD. Finally, as the follow-up information is not included within the NIS database, the postoperative outcomes is not discussed in this study. Despite these limitations, this is the first population-based study that investigated the impact of AF on the outcomes of pancreatic cancer patients undergoing OPD, providing surgeons data and serving as a basis for future prospective studies.